Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method of performing proactive comprehensive geriatric risk screening comprising: receiving at a processing device, individual features data of a patient being assessed for multiple risk types; running, by the processing device, a multi-task predictive model trained to jointly predict multiple target risk types for said individual based on said individual features data and predict a set of risk associations by determining correlations between target risk types, said multi-task predictive model trained based on: data representing risks across multiple vulnerability domains, data representing features of multiple patients, and data representing complete or incomplete observations in risk targets and features of said multiple patients; and trained based on optimizing, at the processing device, a linkage regularization using the features data, the risks across multiple vulnerability domains data and said complete or incomplete observations data, said linkage regularization regulating said multi-task predictive model training, a selecting and ranking of said risk features, and a learning and selecting the set of risk associations, by linking a coefficient matrix relating target features and risk types used in said predictive model and a covariance matrix representing domain knowledge on risk associations; and calculating, by the processing device, a risk score for said jointly predicted multiple target risk types for said individual patient using said trained model and said linkage regularization optimizing; and outputting said patient risk score for each said multiple target risk types in each domain for said individual patient for display via a device providing a user interface; and providing, based on the predicted patient risk score for said multiple target risk types, a course of preventative treatment for the individual patient.
2. The method of claim 1 , wherein said optimizing said linkage regularization comprises: performing an iterative algorithm on said risk feature selection and ranking; applying a thresholding rule to update elements of the covariance matrix representing domain knowledge on risk associations used by the iterative algorithm for the risk feature selection and ranking; and leveraging said performing the iterative algorithm and said applying the thresholding rule.
A method for proactive comprehensive geriatric risk screening uses a processing device to analyze patient data with a multi-task predictive model. This model is trained to jointly predict multiple risk types and their associations, based on patient features, vulnerability domain risks, and observation data. A crucial part of this training involves optimizing a "linkage regularization" technique. This technique regulates the model's training, the selection and ranking of risk-predictive features, and the learning of risk associations by linking a feature-to-risk coefficient matrix with a covariance matrix representing known risk relationships. The optimization of this linkage regularization specifically includes: iteratively selecting and ranking risk features, and applying a thresholding rule to update the elements of the covariance matrix (which contains domain knowledge about risk associations) that are used by this iterative feature selection process. These iterative and thresholding steps are combined to refine the regularization. After training, the method calculates and outputs patient risk scores for display and provides preventative treatment recommendations. ERROR (embedding): Error: Failed to save embedding: Could not find the 'embedding' column of 'patent_claims' in the schema cache
3. The method of claim 2 , wherein said iterative algorithm comprises: running a smoothing proximal gradient algorithm.
Technical Summary: This invention relates to iterative algorithms used in optimization problems, particularly for solving large-scale or complex optimization tasks. The problem addressed is the computational inefficiency and convergence challenges in traditional optimization methods when dealing with non-smooth or constrained objective functions. The invention improves upon prior methods by incorporating a smoothing proximal gradient algorithm within an iterative optimization framework. The iterative algorithm processes an objective function by iteratively updating variables to minimize or maximize the function. The smoothing proximal gradient algorithm is a key component that enhances convergence by approximating non-smooth terms with smooth functions, allowing gradient-based optimization techniques to be applied effectively. This approach helps avoid the computational bottlenecks associated with direct handling of non-smoothness, such as subgradient methods, which can be slow or unstable. The algorithm iteratively refines the solution by applying gradient-based updates while incorporating proximal terms to handle constraints or regularization. The smoothing step ensures that the gradient computations remain stable and efficient, even when the original problem contains non-differentiable components. This method is particularly useful in machine learning, signal processing, and other fields where large-scale optimization is required. The invention provides a more robust and efficient optimization process compared to traditional methods, enabling faster convergence and better handling of complex objective functions.
4. The method of claim 1 , wherein the training further comprises: receiving one or more of expert opinion data, and domain knowledge on risk association data.
This invention relates to a method for training a machine learning model to assess risk associations, particularly in domains where expert knowledge and domain-specific data are critical for accurate predictions. The method addresses the challenge of improving model performance by incorporating structured expert opinions and domain-specific risk association data during training. The core process involves collecting and integrating these additional data sources alongside conventional training data to enhance the model's ability to identify and evaluate risk factors. Expert opinion data may include subjective assessments, historical case analyses, or domain-specific heuristics provided by specialists. Domain knowledge on risk association data encompasses established relationships between variables, causal factors, or risk patterns recognized within the field. By leveraging these inputs, the model refines its understanding of complex risk dynamics, leading to more precise and contextually relevant risk assessments. This approach is particularly valuable in fields such as healthcare, finance, or industrial safety, where domain expertise plays a crucial role in risk evaluation. The method ensures that the model's predictions align with expert consensus and domain-specific insights, reducing reliance solely on statistical patterns from raw data. The integration of these supplementary data sources during training improves the model's robustness and applicability in real-world scenarios where risk assessment requires both data-driven and knowledge-driven inputs.
5. The method of claim 1 , further comprising: determining, by said processor, whether a score of a particular risk target for said individual patient is one of: a high-risk score or low risk score.
This invention relates to a medical risk assessment system that evaluates patient-specific risk factors to determine high-risk or low-risk scores for individual patients. The system processes patient data to identify risk targets, such as disease susceptibility or treatment complications, and assigns a risk score based on predefined criteria. The processor then classifies each risk target as either high-risk or low-risk, enabling personalized medical decision-making. The method involves collecting patient-specific data, analyzing it to identify relevant risk factors, and applying a scoring algorithm to generate a risk assessment. The system may also incorporate additional data sources, such as medical history or genetic information, to refine the risk evaluation. By categorizing risk targets, the invention helps clinicians prioritize interventions and tailor treatment plans to individual patient needs, improving patient outcomes and resource allocation in healthcare settings. The invention addresses the challenge of accurately assessing patient risk in a personalized manner, reducing reliance on generalized risk models and enhancing precision in medical care.
6. The method of claim 1 , wherein the individual features comprise at least one of electronic medical records, answer data from a questionnaire administered to said patient, genetics information, activity data, and diet tracking.
This invention relates to a method for analyzing patient health data to improve medical diagnosis or treatment. The method involves collecting and processing multiple types of patient-specific data to generate insights that assist healthcare providers. The data sources include electronic medical records, patient responses from questionnaires, genetic information, activity tracking data, and diet tracking information. By integrating these diverse data points, the system can identify patterns, correlations, or anomalies that may not be apparent from any single data source alone. This comprehensive approach helps clinicians make more accurate diagnoses, personalize treatment plans, and monitor patient progress over time. The method may also support predictive analytics, such as identifying patients at risk of developing certain conditions based on their combined health data. The system can be implemented in a healthcare setting, such as a hospital or clinic, or as part of a remote monitoring solution for telemedicine. The goal is to enhance patient care by leveraging a holistic view of an individual's health status.
7. An apparatus for performing proactive comprehensive geriatric risk screening, the apparatus comprising: a memory storage device storing a program of instructions; a processor device receiving said program of instructions to configure said processor device to: receive individual features data of a patient being assessed for multiple risk types; run a multi-task predictive model trained to jointly predict multiple target risk types for said individual based on said individual features data and to predict a set of risk associations by determining correlations between target risk types, said multi-task predictive model trained based on: data representing risks across multiple vulnerability domains, data representing features of multiple patients, and data representing complete or incomplete observations in risk targets and features of said multiple patients; and trained based on optimizing linkage regularization using the features data, the received risks across multiple vulnerability domains data and said complete or incomplete observations data, said linkage regularization regulating said multi-task predictive model training, a selecting and ranking of the risk features, and a learning and selecting of the set of risk associations, said linkage regularization linking a coefficient matrix relating target features and risk types used in said predictive model and a covariance matrix representing domain knowledge on risk associations; and calculate a risk score for said jointly predicted multiple target risk types for said individual patient using said trained model and said linkage regularization optimizing; and output said patient risk score for each said multiple target risk types in each domain for said individual patient for display via a device providing a user interface; and provide, based on the predicted patient risk score for said multiple target risk types, a course of preventative treatment for the individual patient.
This invention relates to a system for proactive geriatric risk screening that assesses multiple risk types simultaneously. The system addresses the challenge of identifying and managing various health vulnerabilities in elderly patients, where risks often intersect across different domains. The apparatus includes a processor and memory storing instructions to execute a multi-task predictive model. The model processes individual patient data to predict multiple risk types, such as falls, cognitive decline, or malnutrition, by analyzing correlations between these risks. The model is trained using data from multiple patients, including complete and incomplete observations, to improve accuracy. It employs linkage regularization to optimize training, selecting relevant risk features and associations while incorporating domain knowledge about risk relationships. The system calculates risk scores for each predicted risk type and displays them via a user interface. Additionally, it recommends preventative treatments based on the predicted risks. The approach enhances early intervention by leveraging interconnected risk factors, improving care outcomes for elderly patients.
8. The apparatus of claim 7 , wherein the processor device is further configured to: perform an iterative algorithm on said risk feature selection and ranking; apply a thresholding rule to update elements of the covariance matrix representing domain knowledge on risk associations used by the iterative algorithm for the risk feature selection and ranking; and leveraging said perform the iterative algorithm and said apply the thresholding rule.
This invention relates to a computational apparatus for risk assessment, specifically improving the selection and ranking of risk features in financial or operational domains. The apparatus addresses the challenge of accurately identifying and prioritizing risk factors from large datasets, where traditional methods may fail to capture complex interdependencies or domain-specific knowledge. The apparatus includes a processor configured to execute an iterative algorithm for risk feature selection and ranking. The algorithm iteratively refines feature importance by updating a covariance matrix, which encodes domain knowledge about risk associations. A thresholding rule is applied to adjust elements of this matrix, ensuring that only meaningful risk relationships influence the selection process. This iterative refinement and thresholding step improves the robustness and interpretability of the risk model. The processor further leverages the updated covariance matrix to enhance the iterative algorithm's performance, ensuring that the selected features are both statistically significant and aligned with domain expertise. This approach mitigates overfitting and improves generalization across different risk scenarios. The apparatus is particularly useful in financial risk management, where accurate feature selection is critical for regulatory compliance and decision-making.
9. The apparatus of claim 8 , wherein said iterative algorithm comprises a smoothing proximal gradient algorithm.
This invention relates to an apparatus for processing data using an iterative algorithm, specifically a smoothing proximal gradient algorithm, to solve optimization problems. The apparatus is designed to handle large-scale data processing tasks where traditional optimization methods may be inefficient or computationally expensive. The smoothing proximal gradient algorithm is particularly useful for problems involving non-smooth or non-convex objective functions, which are common in machine learning, signal processing, and statistical modeling. The apparatus includes a processor configured to execute the iterative algorithm, which iteratively updates a solution vector to minimize an objective function. The algorithm incorporates a proximal gradient step to handle non-smooth terms and a smoothing technique to improve convergence. The processor may also include memory to store intermediate results and parameters, ensuring efficient computation. The apparatus may further include input and output interfaces to receive data and output optimized results. The smoothing proximal gradient algorithm is an iterative method that combines gradient descent with a proximal operator to handle non-smooth penalties or constraints. The algorithm first computes a gradient of a smooth component of the objective function, then applies a proximal operator to a non-smooth component, and finally updates the solution with a smoothing step to accelerate convergence. This approach is particularly effective for large-scale problems where exact solutions are computationally infeasible. The apparatus is designed to be flexible, allowing for different problem formulations and constraints, making it suitable for a wide range of applications in data analysis, machine learning, and optimization. The us
10. The apparatus of claim 7 , wherein the processor device is further configured to: receive one or more of expert opinion data, and domain knowledge on risk association data.
This invention relates to an apparatus for processing risk association data, particularly in domains requiring expert analysis. The apparatus includes a processor device configured to receive and analyze various types of input data, including expert opinion data and domain-specific knowledge on risk associations. The processor device is designed to integrate these inputs to assess and quantify risks, likely improving decision-making in fields such as finance, healthcare, or industrial safety. The apparatus may also include a memory device for storing the received data and a communication interface for transmitting processed risk assessments. The processor device further processes the received data to generate risk models or predictions, which can be used to mitigate potential hazards or optimize outcomes. The invention addresses the challenge of incorporating subjective expert insights with structured domain knowledge to enhance risk evaluation accuracy. By automating the integration of these diverse data sources, the apparatus aims to provide more reliable and actionable risk assessments compared to traditional methods that rely solely on manual analysis or limited datasets. The system is adaptable to different industries where risk management is critical, ensuring flexibility in application.
11. The apparatus of claim 7 , wherein the processor device is further configured to determine, whether a score of a particular risk target for said individual patient is one of: a high-risk score or low risk score.
This invention relates to a medical risk assessment apparatus designed to evaluate patient-specific risk levels. The apparatus includes a processor device that analyzes patient data to determine whether a particular risk target for an individual patient falls into one of two categories: a high-risk score or a low-risk score. The processor device is configured to process patient data, which may include medical history, diagnostic results, or other relevant health metrics, to generate a risk score for the patient. The apparatus then classifies this score as either high-risk or low-risk based on predefined thresholds or criteria. This classification helps clinicians identify patients who may require immediate intervention or closer monitoring versus those who are at lower risk. The system may also include a display device to present the risk assessment results to healthcare providers, enabling more informed decision-making. The invention aims to improve patient outcomes by providing a clear, binary risk classification that simplifies risk management in clinical settings. The apparatus may be part of a larger diagnostic or monitoring system, integrating with other medical devices or databases to enhance accuracy and usability.
12. The apparatus of claim 7 , wherein the individual features comprise at least one of electronic medical records, answer data from a questionnaire administered to said patient, genetics, activity data, and diet tracking.
This invention relates to a medical data analysis apparatus designed to improve patient health monitoring and treatment personalization. The apparatus collects and processes diverse patient data to generate actionable insights for healthcare providers. The system integrates multiple data sources, including electronic medical records, patient questionnaire responses, genetic information, activity tracking data, and diet tracking records. By analyzing these inputs, the apparatus identifies patterns, trends, and correlations that inform clinical decisions. The technology addresses the challenge of fragmented patient data by consolidating information from disparate sources into a unified platform. This holistic approach enables more accurate diagnoses, personalized treatment plans, and proactive health management. The apparatus may also incorporate machine learning algorithms to predict patient outcomes or recommend interventions based on the aggregated data. By leveraging comprehensive patient profiles, healthcare providers can deliver more precise and effective care, ultimately improving patient outcomes and operational efficiency in medical settings.
13. A non-transitory computer readable storage medium, tangible embodying a program of instructions executable by the computer for performing proactive comprehensive geriatric risk screening comprising: receiving individual features data of a patient being assessed for multiple risk types; running a multi-task predictive model trained to jointly predict multiple target risk types for said individual based on said individual features data and predict a set of risk associations by determining correlations between target risk types, said multi-task predictive model trained based on: data representing risks across multiple vulnerability domains, data representing features, and data representing complete or incomplete observations in risk targets and features of said multiple patients; and trained based on optimizing linkage regularization using the features data, the risks across multiple vulnerability domains data and said complete or incomplete observations data, said linkage regularization regulating said multi-task predictive model training, a selecting and ranking of the risk features, and a learning and selecting of the set of risk associations, said linkage regularization linking a coefficient matrix relating target features and risk types used in said predictive model and a covariance matrix representing domain knowledge on risk associations; and calculating a risk score for said jointly predicted multiple target risk types for said individual patient using said trained model and said linkage regularization optimizing; and outputting a patient risk score for each said multiple target risk types in each domain for said individual patient for display via a device providing a user interface; and providing, based on the predicted patient risk score for said multiple target risk types, a course of preventative treatment for the individual patient.
This invention relates to a computer-implemented system for proactive geriatric risk screening, addressing the challenge of assessing multiple health risks in elderly patients simultaneously. The system uses a multi-task predictive model trained to jointly predict various risk types (e.g., falls, cognitive decline, malnutrition) based on patient features such as medical history, demographics, and lifestyle factors. The model also identifies correlations between different risk types, improving accuracy by leveraging shared vulnerabilities across domains. The model is trained using data from multiple patients, including complete and incomplete observations, and incorporates linkage regularization to optimize feature selection, risk association learning, and model training. Linkage regularization connects a coefficient matrix (relating patient features to risk types) with a covariance matrix (encoding domain knowledge about risk associations), ensuring predictions align with medical expertise. The system calculates risk scores for each target risk type and displays them via a user interface, along with personalized preventative treatment recommendations. By integrating multiple risk assessments into a single model, this approach enhances early intervention for geriatric patients, reducing the burden of separate evaluations and improving care coordination. The system’s ability to handle incomplete data and learn risk associations makes it robust for real-world clinical applications.
14. The non-transitory computer readable storage medium of claim 13 , wherein optimizing said linkage regularization comprises: performing an iterative algorithm on said feature selection and ranking; applying a thresholding rule to update elements of the covariance matrix representing domain knowledge on risk associations used by the iterative algorithm for the risk feature selection and ranking; and leveraging said performing the iterative algorithm and said applying the thresholding rule.
This invention relates to machine learning techniques for optimizing feature selection and ranking in risk assessment models, particularly in domains where domain knowledge about risk associations is critical. The problem addressed is the challenge of effectively incorporating prior domain knowledge into automated feature selection processes to improve the accuracy and interpretability of risk prediction models. The invention describes a method for optimizing linkage regularization in feature selection and ranking algorithms. This involves performing an iterative algorithm to evaluate and rank features based on their relevance to risk prediction. During this process, a thresholding rule is applied to update elements of a covariance matrix, which encodes domain knowledge about risk associations. The covariance matrix guides the iterative algorithm by emphasizing or de-emphasizing certain feature relationships based on known risk factors. The iterative algorithm and thresholding rule work together to refine feature selection, ensuring that the final model incorporates both data-driven insights and domain expertise. The approach is particularly useful in applications such as healthcare, finance, or any field where risk assessment must balance statistical evidence with established domain knowledge. By dynamically adjusting the covariance matrix, the method ensures that the selected features align with known risk associations, leading to more reliable and interpretable risk models. The invention improves upon traditional feature selection methods by integrating domain knowledge in a structured, iterative manner.
15. The non-transitory computer readable storage medium of claim 14 , wherein said iterative algorithm comprises a smoothing proximal gradient algorithm.
A non-transitory computer-readable storage medium stores instructions for a machine learning system that processes data using an iterative algorithm. The system includes a processor and memory, where the memory stores training data and a model. The iterative algorithm is designed to optimize a loss function by iteratively updating model parameters. The algorithm includes a proximal gradient step that applies a gradient-based update to the model parameters and a proximal operator that enforces constraints or regularization. The proximal gradient step computes gradients of the loss function with respect to the model parameters and updates the parameters by moving them in the direction of the negative gradient. The proximal operator then modifies the updated parameters to satisfy predefined constraints, such as sparsity or boundedness. The iterative process repeats until convergence or a stopping criterion is met. The smoothing proximal gradient algorithm is a specific type of proximal gradient method that incorporates smoothing techniques to improve convergence speed or stability. This approach is particularly useful in machine learning tasks where the loss function or constraints are non-smooth, such as in sparse regression or support vector machines. The system efficiently handles large-scale data by leveraging iterative optimization techniques that balance computational efficiency and solution accuracy.
16. The non-transitory computer readable storage medium of claim 13 , wherein the training further comprises receiving one or more of expert opinion data, and domain knowledge on risk association data.
The invention relates to a computer-implemented system for training machine learning models to assess risk associations in data. The system addresses the challenge of accurately identifying and quantifying risks in complex datasets by incorporating diverse sources of information to improve model performance. The training process involves processing structured and unstructured data to extract relevant risk indicators. Additionally, the system integrates expert opinion data and domain-specific knowledge on risk associations to enhance the model's ability to recognize patterns and relationships that may not be apparent from raw data alone. This approach ensures that the model's predictions are grounded in both empirical evidence and human expertise, leading to more reliable risk assessments. The trained model can then be applied to new data to predict risk levels, providing actionable insights for decision-making in fields such as finance, healthcare, and cybersecurity. By leveraging multiple data sources and expert knowledge, the system aims to overcome the limitations of traditional risk assessment methods that rely solely on statistical analysis or predefined rules.
17. The non-transitory computer readable storage medium of claim 13 , further comprising: determining whether a score of a particular risk target for said individual patient is one of: a high-risk score or low risk score.
This invention relates to a computer-implemented system for assessing patient risk, particularly in medical or healthcare applications. The system evaluates patient data to determine risk levels, such as high-risk or low-risk scores, for specific risk targets. The risk assessment is based on analyzing patient-specific data, which may include medical history, test results, or other clinical indicators. The system processes this data using predefined criteria or algorithms to generate a risk score for each patient. The risk score is then classified as either high-risk or low-risk, enabling healthcare providers to prioritize interventions or treatments accordingly. The system may also include additional features, such as generating alerts or recommendations based on the risk classification. The invention aims to improve patient outcomes by providing data-driven risk assessments that guide clinical decision-making. The risk evaluation process is automated, reducing manual effort and potential human error in risk determination. The system is designed to be integrated into existing healthcare information systems, ensuring seamless access to patient data and risk assessments. The invention addresses the need for accurate, efficient, and personalized risk stratification in healthcare settings.
Unknown
January 14, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.